0

Racial Disparity in Primary Hepatocellular

Carcinoma: Tumor Stage at Presentation,
Surgical Treatment and Survival
Dana Sloane, MD; Hegang Chen, PhD; and Charles Howell, MD
Baltimore, Maryland

Financial support: This work was supported by Public Health

Service grant #1 K24 DK072036-01 (Howell) and MO1 RR16500.
Objectives: The incidence and mortality rates from primary
hepatocellular carcinoma (HCC) are higher in black Americans compared to whites. The goal of this study was to
determine if there are racial disparties in HCC stage at
diagnosis and treatment.
Methods: We compared patient age, tumor stage, rates of
surgical intervention and survival in black (n=1,718) and
white (n=9,752) HCC cases between 1992 and 2001 in the
Surveillance, Epidemiology and End Results (SEER)-1 1 + Alaska database.
Results: Black HCC cases were significantly younger at diagnosis (p<0.0001). Black cases were more likely to have
regional and distant metastasis at presentation (p<0.0005)
and were less likely to have surgery performed (p<0.001).
The racial difference in surgery treatment was significant
among patients with localized (p=0.001) and regional (p=
0.012) HCC, but not with distant HCC. Overall survival rates
were lower in blacks (p=0.0033). Survival was similar in blacks
and whites with regional and distant disease. Yet, among
patients with localized HCC, survival rates were lower in
blacks (p=0.0030).
Conclusions: Black HCC patients have more advanced
tumor stage at diagnosis and lower rates of both surgical
intervention and survival. The racial disparities in surgical
treatment utilization and survival were most strking between
black and white HOC patients with localized HOC.

Key words: hepatocellular carcinoma * racial dispanty

2006. From the University of Maryland School of Medicine, Baltimore MD.
Send correspondence and reprint requests for J NatI Med Assoc. 2006;98:
1934-1939 to: Dr. Dana A. Sloane; phone: (202) 877-7108; e-mail:
dana.sloane@medstar.net

1934 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION

INTRODUCTION
Primary liver cancer is the fifth most common
malignancy in the world and accounts for up to 1
million deaths per year worldwide. The incidence
of hepatocellular carcinoma (HCC) has been rising in
the United States over the past 20 years.1'2 The ageadjusted incidence for HCC in the United States
increased from 1.4 per 100,000 in 1975-1977 to 3.0 per
100,000 in 1996-1998. There was a concurrent rise in
HCC mortality from 1.7 per 100,000 to 2.4 per 100,000
during 1991-1995. The increased incidence of HCC has
been attributed, in part, to an increase in the prevalence
of end-stage liver disease due to chronic hepatitis-B and
-C infections.3'4 Both the incidence and mortality rates
for HCC are 2-3 times higher in black Americans than
in whites." 2 HBV, HCV, concurrent HBV and HCV, and
diabetes mellitus plus viral hepatitis were significantly
more common in black HCC cases compared to white
cases in the Nationwide Inpatient Sample for the year
2000.5 Thus, differences in the prevalence of HCC risk
factors may account for the racial differences in HCC
incidence. However, the explanation for the higher HCC
mortality rates in blacks is unknown.
Previous studies have found more advanced tumor
stage at diagnosis and lower rates of surgery with curative
intent for colorectal, breast, cervical and prostate cancers
in black Americans relative to white Americans.6 To better define the racial disparity in HCC in the United States,
we compared the HCC tumor stage at diagnosis, utilization of surgical treatment, and survival in black and white
HCC cases in the United States for 1992-2001.

METHODS
Description of Database
The National Cancer Institute's Surveillance, Epidemiology and End Results (SEER) program is a comprehensive source of population-based information on
cancer incidence and survival in the United States.7 We
utilized data from the SEER-lI + Alaska database registry from 1992-2001. The registry includes data on
VOL. 98, NO. 12, DECEMBER 2006

PRIMARY HEPATOCELLULAR CARCINOMA

cancer cases from Atlanta, GA; Connecticut' Detroit,

MI; Hawaii; Iowa; New Mexico; Los Angeles, San JoseMonterey and San Francisco-Oakland, CA; SeattlePuget Sound, WA; and Utah. Starting in 2001, cases
from Louisiana, Kentucky, New Jersey and the remaining areas of California were also included.

Patient Characteristics and

Clinical Variables
The intended cohort was all cases in the SEER database
with a diagnosis ofprimary HCC, using ICD-9 code 155.0
(cases through 1998), and ICD-10 code C22.0 (cases from
1999 and beyond). HCC cases with race classified as either
white or black were selected for this study. The following
patient variables were examined: age at diagnosis, gender,
tumor stage, utilization of surgical therapy and status (dead
or alive). In the SEER database, tumor stage is categorized
as localized, regional spread or distant spread. Localized
cancer is defined as disease that is limited to the organ in
which it began, without evidence of spread. Regional
spread is cancer that has spread beyond the original (primary) site to nearby lymph nodes, organs and tissues. Distant spread is cancer that has extended beyond the primary
site to distant organs or lymph nodes.
SEER classifies surgical interventions into several
categories: surgery performed; surgery recommended,
unknown if performed; surgery refused; surgery not recommended; surgery contraindicated; unknown intervention; and unknown-death certificate only. The mortality data reported by SEER are obtained from the
National Center for Health Statistics.

Data Analysis
The age-adjusted HCC incidence for black and white
cases was determined using U.S. population data provided by SEER, and compared using the Kruskal-Wallis test.
The Chi square with post hoc cell analysis, and observed
versus expected cases were used to compare the age distributions of HCC, tumor stages and surgical treatment
utilization rates between black and white cases.
Survival by race, HCC stage, age and gender were
estimated using the Kaplan-Meier method and compared using the log-rank test. Cox regression models
were used to compare HCC survival by race, adjusting
for age, gender, surgical procedure and HCC stage.
Two-sided p values of <0.05 were considered statistically significant. Analyses were performed using SAS version 9.1 (SAS Institute, Cary, NC). Cases with status
"alive were deemed to be uncensored, whereas those
with "dead" status were deemed to be censored. Survival times were adjusted for age.

RESULTS
We identified 9,752 white and 1,718 black HCC cases
in the SEER database between 1992 and 2001. Males
accounted for 68% of both the white and black cases. The
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION

age-adjusted HCC incidence was higher in blacks than in

whites (4.6 per 100,000 vs. 2.6 per 100,000, p<0.0001). In
addition, both black and white males had a greater HCC
incidence than black and white females (7.6 per 100,000 vs.
2.3 per 100,000 and 4.1 per 100,000 vs. 1.4 per 100,000,
respectively; p<0.0001). The age distribution of HCC cases
among blacks and whites is shown in Figure 1. HCC was
uncommon in cases <40 years of age in both races. Compared to white cases, a greater proportion of black cases
were <70 years of age, and a smaller proportion were >70
(p<0.0001). The racial difference in age distribution applied
to both male (p<0.001) and female cases (p=0.02).
Tumor stages at diagnosis (Figure 2) were significantly different between black and white HCC cases
overall [p<0.0005 (Chi square)] as well as within each
gender (p<0.001 for males; p=0.02 for females). Black
patients were more likely to have regional (27.5% vs.
22.4%) and distant (22.7 vs. 19.9%) HCC, and less likely to have localized (27% vs. 30.1%, Figure 2) and
unstaged disease (22.8% vs. 27.7%). White males were
more likely to have regional and distant HCC and less
likely to have localized tumor than white females
(p<0.0001). There was no difference, however, in HCC
stage between black males and females (p=0. 14).

Surgical Treatment Utilization and

Patient Survival
There was a significant association between tumor
stage and surgical intervention (p<0.0001) (Figure 3).
Figure 1. Age distribuflon of hepatocellular
carcinoma cases
* Black (n=1,718)

20LI White (n=9,752)

15-

O
101

p<O.0001

5-

0-

~~o~~~or-O

O O O

-.
-

A$$b8$u,A,,L,,,A
N
Ag
e
0
_
N ee

co

Age

VOL. 98, NO. 12, DECEMBER 2006 1935

PRIMARY HEPATOCELLULAR CARCINOMA

Surgical treatment was "not recommended" more often

for patients with regional and distant HCC spread than
for patients with localized and unstaged HCC (p<0.001).
Figure 2. HCC tumor stage
* Black
35

White

p<O.OOOl

3025-

20
15-

10
5-

0*

c~~~~

o
-J~~~04

Figure 3. Surgical treatment by HCC stage

* Localized EZ lBgional

P71

Distant "- Unstaged

6050

P<0.0001

40
C:

330

302010
0

1936 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION

Conversely, surgery was performed more often for

patients with localized HCC and less often for patients
with regional, distant and unstaged HCC. There were also
significant differences in other surgical intervention categories based on HCC stage. Cases with distant spread had
contraindications more often than cases with localized
and regional spread, whereas the unstaged cases were less
likely to have contraindications. Unstaged HCC cases
tended to refuse surgical intervention more often and
were more likely to be categorized as "intervention
unknown" and "unknown-death certificate." There were
no differences by stage in the percentage of cases classified as "surgery recommended but unknown if done."
There was also a significant difference in surgical
treatment between black and white patients (p<0.0001).
Assuming race and surgery were independent variables,
795 blacks were expected to have "surgery not recommended," and 223 were expected to have "surgery performed." The observed numbers were 898 and 160,
respectively. A similar percentage of blacks and whites
were classified as: surgery contraindicated (7.4% vs.
6.6%); surgery refused (1.6% vs. 1.5%); surgery recommended, with unknown outcome (0.5% each); unknown
intervention (26.3% vs. 27.5%); and, unknown-death
certificate only (3.5% vs. 4.8%). Analysis of post hoc
cell contributions indicated that the racial disparity
explained the "surgery not recommended" and "surgery
performed" categories. Thus, when the Chi-squared
analysis was restricted to these two categories, black
patients were still significantly more likely to have "surgery not recommended" and less likely to have "surgery
performed" than white patients (p<0.0001). A comparison of the proportions of "surgery not recommended"
and "surgery performed" by race and HCC stage
showed racial differences only in localized (p=0.001)
and regional (p=0.012) HCC cases (Figure 4). blacks
with localized (Figure 4A) and regional (Figure 4B)
spread were more likely to have "surgery not recommended" and, subsequently, were less likely to have surgery performed. There were no differences in surgical
intervention between black and white cases with distant
(Figure 4C) and unstaged (Figure 4A) disease.
Black cases had significantly lower overall survival
(p=0.0033) (Figure 5A). Survival among black male
cases was significantly lower than for white male cases
(p=0.02). There was no difference in survival between
black and white female HCC cases. When compared by
age category, the racial disparity in survival was statistically significant only for ages 20-45 (p<0.0001) and
46-69 (p=0.0012) years. Survival in both races was also
related to HCC stage, with the longest survival in localized HCC cases followed by regional, distant and
unstaged cases (p<0.0001) (Figure 5B). Comparison of
survival by race and tumor stage demonstrated significantly lower survival in blacks compared with whites
with localized HCC (p=0.003; Figure 6A); there was no
VOL. 98, NO. 12, DECEMBER 2006

PRIMARY HEPATOCELLULAR CARCINOMA

difference in survival between black and white cases

with regional (Figure 6B), distant (Figure 6C) and
unstaged HCC (Figure 6D). Survival among black HCC
cases was lower [p=0.002; 95% hazard ratio confidence
interval (CI) 1.041-1.20] after adjusting for age, gender,
tumor stage and surgical treatment intervention using
Cox regression analysis.

DISCUSSION
Black HCC cases in our study tended to be younger
and to have more advanced tumor stage. In addition,
black cases were less likely to have surgery recommended and, consequently, less likely to have surgery performed. However, the lower surgical and patient survival rates among black patients were not explained
solely by racial differences in tumor stage.
Previous studies have reported more-advanced
tumor stages and lower rates of surgery with curative
intent in black Americans with colorectal, breast, lung
and head/neck cancers relative to white Americans.68-10
In an analysis of SEER data from 1973-1998, Shavers
and Brown found lower five-year survival rates for
African Americans than for whites for all the major cancer sites.9 In addition, their study revealed racial disparities in cancer treatment, including treatment with primary and adjuvant chemotherapy, as well as surgery
with curative intent. Our study found similar disparities
between black and white HCC patients in tumor stage at
presentation, surgical treatment and survival.
The more advanced HCC stage at diagnosis may certainly account for the lower rate of surgical treatment
with curative intent and for lower survival in black
patients. A clinical decision against either surgical
resection or liver transplantation is appropriate for many
patients with localized HCC and all patients with
regional and distant HCC. However, several findings in
the current study suggest that the racial disparities in
surgical treatment and patient survival are not completely explained by differences in tumor stage at diagnosis.
First, the most striking disparities in surgical treatment
and survival were found in patients with HCC localized
to the liver. Providers tended to recommend against surgery more often-and surgery was thus performed less
often-in blacks with localized and regional HCC
spread. In addition, blacks with localized HCC had a
significantly lower overall survival time. Patients of
both races who had surgery had significantly longer survival than patients for whom surgery was not recommended. Yet, among the patients who survived, black
patients were recommended for surgical therapy 50%
less often than their white counterparts.
Issues of socioeconomic disparity and racial differences in access to care may certainly play a role in these
observations. However, SEER does not contain data pertaminin to income or insurance status. Other investigators
have identified racial inequities in access to care, average
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION

Figure 4. Surgical intervention by stage

EL

U Black

White

A. Localized HCC

50

p<0.001I
40 -

20
.0

20

Not recommended

60120.1

Performed

Surgery

B. Regional spread

7060

~~~~~p<0.0 12

50-

~40
20-

0
Not recommended

Performed

Surgery

VOL. 98, NO. 12, DECEMBER 2006 1937

PRIMARY HEPATOCELLULAR CARCINOMA

unequal access to healthcare for

HCC screening and treatment,
and
delayed referral for surgery.
A. By race (p=0.0033)
This study has a few limita1.00
tions. The SEER database
includes only certain geographic
0r
regions. Thus, this cohort might
0.76
IL
not be completely representative
c
of national black and white HCC
0
b.60
cases. Second, decisions concerning surgical treatment for HCC
localized to the liver typically take
~026
into consideration the size of the
tumor, the number of tumor nod00.00
ules and the distribution of HCC
inside the liver. Thus, the broader
0
20
40 ~60
80
100
120
140
criteria for localized tumor stage
Survival Tif. (Morths)
used by the SEER program could
Whit
STRATA:
Black
obscure potential differences in
staging that would have been
B. By HCC stage (p<0.0001)
identified using more convention1.00
al criteria, such as the TMN,

Milan or Okuda system. Third,

0
because SEER includes all-cause
~0.76
mortality, it is not possible to
determine HCC-related deaths
from other causes for mortality.
20.60
\
Moreover, as previously mentioned, we were not able to exam~026
ine other factors that might
influence the tumor stage at presentation, such as racial differences
0.001
-------------_____________________
IIII
IIII
in healthcare payers, health-seek0 20
40
60
8
100
120
140
behavior and tumor behavior.
ing
Survival Mme (Months)
Given our findings of racial
StagooDletant
I S -Stag..R.glonal
STRATA:
b
disparity in surgical treatment and
-StagooLocalized
I -II -StbgUnstag.d
survival for HCC patients, we
suggest that future studies address
household income and insurance status. Reid et al. found the importance of healthcare system, healthcare provider
that blacks were underrepresented on the United Network and patient factors that lead to racial disparities in HCC
for Organ Sharing (UNOS) liver transplant waiting list, treatment utilization.
compared to the proportion of blacks in the U.S. population."1 That study also revealed that blacks have a higher ACKNOWLEDGEMENTS
prevalence of chronic viral hepatitis and higher mortality
The authors would like to thank Olga Goloubeva,
rates from liver disease. Data on household income and PhD, MSc for her contributions to the statistical analysis
insurance status were not available in the SEER database. in this study.
However, Yu et al. used data from the 2000 Nationwide
Inpatient Sample to demonstrate that blacks with HCC REFERENCES
1. El-Serag HB, Mason AC. Rising incidence of hepatocellular carcinoma in
were more likely to have an annual household income the
United States. N Engl J Med. 1999;340:745-750.
<$25,000 and were less likely to have Medicare and pri- 2. El-Serag
Davila JA, Petersen NJ, et al. The continuing increase in the
vate healthcare payers.5 In addition, Guidry et al. reported incidence ofHB,hepatocellular
carcinoma in the United States: an update.
that black Americans tended to reach their insurance Ann Intem Med. 2003;1 39:817-823.
spending limits more frequently than whites and were 3. Di Bisceglie AM, Lyra AC, Schwartz M, et al. Hepatitis C-related hepatocellular carcinoma in the United States: influence of ethnic status. Am J
more likely to lose medical insurance coverage following
Gastroenterol. 2003;98:2060-2063.
a cancer diagnosis.'2 Thus, differences in healthcare pay4. Davila JA, Morgan RO, Shaib Y, et al. Hepatitis C infection and the
ers between blacks and whites could contribute to
increasing incidence of hepatocellular carcinoma: a population-based
Figure 5. Patient survival

14

1938 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION

VOL. 98, NO. 12, DECEMBER 2006

PRIMARY HEPATOCELLULAR CARCINOMA

Figure 6. HCC survival by race and stage

B. By race and regional stage (p=0.0030)

A. By race and localized stage (p=0.2321)

lc1.00
Ui.

I 1.008

0.75.

0.75.

U.
0

~~~~~~~~~~~~~0.50

0.60.

---.--

--

I.0, 0

0.0

20

40

60

100

80

120

140

20

STRATA:

Black

---

40

100

80

60

120

140

Survival Time (Months)

Survival Thme (Months)

-

Black

STRATA:

White

--

- -

White

D. By race and unstaged disease (p=0.2452)

C. By race and distant stage (p=O.1 922)

1.00.

1.00.
C

C0
~ ~~~~
~
~~~~ 0.75

~~0.76.

~~~~~~~~~~

~~~~~~~~~~~~~~~~~~~~~~~~U.

U.
C

0~ ~ ~ ~ ~ ~ o

~025-

,026

K.

00.001
0

20
STRATA:

80
80
100
40
Survival Time (Months)
--White
Black

120

140

study. Gastroenterology. 2004;1 27:1372-1380.

5. Yu L, Sloane DA, Guo C, et al. Risk factors for prmary hepatocellular carcinoma in Black and White Americans in 2000. Clin Gastroenterol Hepatol.
2006;4:355-360.
6. Ward E, Jemal A, Cokkinides V, et al. Cancer disparities by race/ethnicity
and socioeconomic status. CA CancerJ Clin. 2004;54:78-93.
7. National Cancer Institute D, Surveillance Research Program, Cancer Statistics Branch. Surveillance, Epidemiology, and End Results (SEER) Program
(www.seer.cancer.gov). SEER*Stat Database: Incidence - SEER 11 Regs +
AK Public-Use, November 2003 Sub (1973-2001 varying); 2004.
8. Lathan CS, Neville BA, Earle CC. The effect of race on invasive staging
and surgery in non-small-cell lung cancer. J Clin Oncol. 2006;24:413-418.
Epub 12/19/05.
9. Shavers VL, Brown ML. Racial and ethnic disparties in the receipt of cancer treatment. J Natl Cancer Inst. 2002;94:334-357.
10. Tomar SL, Loree M, Logan H. Racial differences in oral and pharyngeal
cancer treatment and survival in Florida. Cancer Causes Control. 2004; 15:
601-609.
11. Reid AE, Resnick M, Chang Y, et al. Disparity in use of orthotopic liver
transplantation among blacks and whites. Liver Transpl. 2004;10:834-841.
12. Guidry JJ, Aday LA, Zhang D, et al. Cost considerations as potential
barriers to cancer treatment. Cancer Pract. 1998;6:182-187. A

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION

20
STRATA:

80
120
100
60
Survival lime (Months)
----White
Black

40

140

REUSE THIS
CONTENT
To photocopy, e-mail, post on Internet or
distribute this or any part of JNMA, please
visit www.copyright.com.

We Welcome Your Comments

The Journal of the National Medical Association
welcomes your Letters to the Editor about
articles that appear in the JNMA or issues
relevant to minority healthcare. Address
correspondence to EditorJNMA@nmanet.org.

VOL. 98, NO. 12, DECEMBER 2006 1939